Datasets:
File size: 2,862 Bytes
a785ccd 8685f4e a785ccd 8685f4e a785ccd 8685f4e a785ccd 8685f4e a785ccd 8685f4e a785ccd | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 | """Assemble the public Hugging Face dataset folder for Skills Desk.
Writes OUT/tasks.jsonl (one row per task id: training or reserved split, family, skill, domain, title,
output path and the fixed instance's instruction), OUT/manifest/ (training and reserved id lists) and
OUT/source/ (environment, tools, tests, Dockerfile). The README card is kept by hand. No screening
numbers are written here; v2's training selection and its screen counts are in manifest/ (v2-selection.json).
Reserved results are a baseline, not a label.
Usage: PYTHONPATH=. .venv/bin/python tools/make_dataset.py --out ../outputs/skills-desk-dataset
"""
import argparse
import json
import os
import random
import shutil
import sys
sys.path.insert(0, ".")
from terminal_desk.families import ALL_FAMILIES, instance_seed, parse_task_id, reserved_task_ids # noqa: E402
SOURCE = ["terminal_desk", "tools", "tests", "manifest", "Dockerfile", "openenv.yaml", "README.md", "LICENSE",
"NOTICE", ".dockerignore", "vendor"]
SKILL = {"ev": "evidence", "cmp": "compute", "tx": "transform", "fix": "repair", "plan": "constraints"}
def skill_of(fid: str) -> str:
return SKILL[fid.removeprefix("res_").split("_")[0]]
def main() -> None:
p = argparse.ArgumentParser()
p.add_argument("--out", required=True)
p.add_argument("--manifest", default="manifest/training_tasks_v2.txt")
a = p.parse_args()
train = [ln.strip() for ln in open(a.manifest) if ln.strip() and not ln.startswith("#")]
os.makedirs(a.out, exist_ok=True)
with open(os.path.join(a.out, "tasks.jsonl"), "w") as fh:
for split, ids in (("train", train), ("reserved", reserved_task_ids())):
for tid in ids:
fid, level, _ = parse_task_id(tid)
fam = ALL_FAMILIES[fid]
built = fam.build(level, random.Random(instance_seed(tid))) # the fixed instance reset serves
fh.write(json.dumps({
"task_id": tid, "split": split, "family": fid, "skill": skill_of(fid), "domain": fam.domain,
"level": level, "title": fam.title, "output": "/app/" + built.artifact.path,
"instruction": built.instruction,
}) + "\n")
shutil.rmtree(os.path.join(a.out, "manifest"), ignore_errors=True)
shutil.copytree("manifest", os.path.join(a.out, "manifest"))
for name in ("LICENSE", "NOTICE"):
shutil.copy(name, os.path.join(a.out, name))
src = os.path.join(a.out, "source")
shutil.rmtree(src, ignore_errors=True)
for name in SOURCE:
if os.path.isdir(name):
shutil.copytree(name, os.path.join(src, name), ignore=shutil.ignore_patterns("__pycache__", "qwen_tok"))
else:
shutil.copy(name, os.path.join(src, name))
print("dataset folder ready:", a.out)
if __name__ == "__main__":
main()
|